Torque Ripple Minimization in a Switched Reluctance Drive

Torque Ripple Minimization in a Switched Reluctance Drive by Neuro-fuzzy Compensation Luís O. A. P. Henriques Luís G. B. Rolim Walter I. Suemitsu Paulo J. C. Branco Joaquim A. Dente Federal University of Rio de Janeiro COPPE Brasil Mechatronics Laboratory Instituto Superior Técnico (IST) Portugal èAdvantages u. High efficiency u. Low manufacturing cost u. Fault tolerant u. Reliable u. Easy to repair èDisadvantages u. Torque ripple u. Nonlinear model èApplications u. Traction u. Heavy-duty applications u. Home appliances Diagram of proposed SR torque ripple compensation scheme Compensated Current pulses after 10 learning iterations Switched Reluctance Motor - Complete Simulated System Torque signal without compensation (constant current and velocity) Compensated Torque Harmonics Triangular fuzzy sets Conclusions èNeuro-fuzzy compensating mechanism to ripple reduction was investigated èCompensating signal added in current waveform was used to minimize the torque ripple èBell shape function produces better ripple reduction in all harmonic content èFuture investigation: Compensated Torque Harmonics Bell fuzzy sets u. Application of this concept in an experimental drive u. Incorporate another signal to be trained Torque signal with compensation (constant current and velocity) after 10 learning iterations Compensated Torque Harmonics Gaussian fuzzy sets Support